Our project welcomes external contributions. If you have an itch, please feel free to scratch it.
For more details on the contributing guidelines head to the Docling Project community repository.
We use uv as package and project manager.
To install uv, check the documentation on Installing uv.
You can use the uv sync to create a project virtual environment (if it does not already exist) and sync
the project's dependencies with the environment.
uv syncIf you need to work with a specific version of Python, you can create a new virtual environment for that version and run the sync command:
uv venv --python 3.12
uv syncMore detailed options are described on the Using Python environments documentation.
Simply use the uv add command. The pyproject.toml and uv.lock files will be updated.
uv add [OPTIONS] <PACKAGES|--requirements <REQUIREMENTS>>We use the following tools to enforce code style:
A set of styling checks, as well as regression tests, are defined and managed through the pre-commit framework.
To ensure that those scripts run automatically before a commit is finalized, install pre-commit on your local repository:
pre-commit installTo run the checks on-demand, run:
pre-commit run --all-filesNote: Checks like Ruff will "fail" if they modify files. This is because pre-commit doesn't like to see files modified by its hooks. In these cases, git add the modified files and git commit again.
When submitting a new feature or fix, please consider adding a short test for it.
When a change improves the conversion results, multiple reference documents must be regenerated and reviewed.
The reference data can be regenerated with
DOCLING_GEN_TEST_DATA=1 uv run pytestAll PRs modifying the reference test data require a double review to guarantee we don't miss edge cases.
We use MkDocs to write documentation.
To run the documentation server, run:
mkdocs serveThe server will be available at http://localhost:8000.
Run the following:
mkdocs gh-deploy